A physical domain-based substructuring as a framework for dynamic modeling and reanalysis of systems

A comprehensive physical domain-based formulation of reduced-order models based on dominant and residual normal modes and interface reduction is presented. The dynamic behavior of the substructures is characterized by the dominant fixed interface normal modes and by the static contribution of higher...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Computer methods in applied mechanics and engineering 2017-11, Vol.326, p.656-678
Hauptverfasser: Jensen, H.A., Araya, V.A., Muñoz, A.D., Valdebenito, M.A.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:A comprehensive physical domain-based formulation of reduced-order models based on dominant and residual normal modes and interface reduction is presented. The dynamic behavior of the substructures is characterized by the dominant fixed interface normal modes and by the static contribution of higher order modes. Interface reduction is accomplished by using a reduced number of interface modes. Special attention is considered to the proper treatment of residual normal modes in the context of system reanalyses and sensitivity analyses. The efficiency of the resultant formulation is evaluated in the framework of dynamic response characterization, modal sensitivity analysis and uncertainty propagation analysis. The effectiveness of the proposed model reduction technique is demonstrated by means of numerical examples involving two structural models. Numerical results show that the technique allows an effective dynamic modeling and reanalysis of a class of structural models. •Accurate reduced-order model is obtained.•Approach allows an efficient dynamic modeling of structural systems.•Interface reduction is achieved by using few characteristic constraint modes.•Modal sensitivity analysis is performed very efficiently.•Time history and frequency responses are predicted with high accuracy.
ISSN:0045-7825
1879-2138
DOI:10.1016/j.cma.2017.08.044